pgd node start#

Synopsis#

The pgd node start command starts the Postgres instance on a node, using pg_ctl directly. It doesn’t require systemd or another service manager to be configured on the node.

Users and roles#

Runs pg_ctl start as an OS-level operation and doesn’t check any PGD or Postgres role. Run the command as the OS user that owns the node’s data directory. See User roles .

Syntax#

pgd node <NODE_NAME> start [OPTIONS]

Where <NODE_NAME> is the name of the node to start.

Options#

Short

Long

Description

-D

--pgdata <PG_DATA>

The data directory of the node. Required. Also set with the PGDATA environment variable.

--bindir <BINDIR>

Path to the Postgres bin directory. Defaults to the directory where pgd resides.

--dbname <DBNAME>

Database to connect to for node-name verification. Defaults to pgddb for PGD nodes, edb for EPAS, or postgres otherwise.

-l

--log <FILENAME>

Append the server log to FILENAME. Defaults to postgres-<port>.log in the current directory.

-W

--no-wait

Don't wait until the operation completes.

-t

--timeout <SECS>

Seconds to wait when not using -W. Defaults to 60. Also set with the PGCTLTIMEOUT environment variable.

-s

--silent

Only print errors, no informational messages. Can't be used with --verbose.

-O

--options <OPTS>

Options to pass through to the postgres command. Repeatable.

--verbose

Verbose output. Can't be used with --silent.

See also Global Options .

Examples#

Start the Postgres instance on a node#

pgd node kaboom start --pgdata /var/lib/edb-pge/17/main

Start a node without waiting for completion#

pgd node kaboom start --pgdata /var/lib/edb-pge/17/main --no-wait